Marc MAGRO
Integrability in Gauge and String Theory 2012 - ETH Zürich
1204.0766, 1204.2531, 1206.6050
With F. DELDUC (ENS Lyon) and B. VICEDO (York / Hertfordshire)
Plan
Part I: - Motivation- Non-ultralocality- Difficulties
Part II: Review of first steps of Faddeev-Reshetikhin procedure
Part III: Generalization to the superstring- Alleviation of non-ultralocality- Link with Pohlmeyer reduction
Part IV: Remarks and conclusion
Motivation
Construct corresponding Quantum Integrable Lattice Model
- Long-term goal...- Would provide a proof of quantum integrability
Motivation & Difficulties
Construct corresponding Quantum Integrable Lattice Model
- Necessary Step = Classical integrable discretization
Already very difficult !
Old problem for integrable Sigma Models !
Difficulty comes from their Non-ultralocality
Review: - What is Non-ultralocality - Why it is the origin of difficulties
- Lattice
N sites with periodic conditions
Classical integrable discretization
[Freidel-Maillet '91]Freidel-Maillet Quadratic Algebra
Freidel-Maillet Quadratic Algebra
Antisymmetry
Freidel-Maillet Quadratic Algebra
Jacobi identity
Monodromy has Poisson bracket:
Integrability
Freidel-Maillet Quadratic Algebra
Non-ultralocality
Consider b=0: - Previous conditions imply c=b=0 and a=d with
a solution of modified classical Yang-Baxter equation
- Corresponds to ultralocal model and in particular
Comes from b and c
Continuum Limit
- PB of Lax matrix are of the r/s form with r and s stemming from (a,b,c,d)- Non-ultralocality carried by the matrix s [Maillet '85 '86]
Spatial Lax Matrix
1. Start from continuum2. Hamiltonian Lax matrix known 3. Its PB are of the r/s form
First difficulty for the superstring
[M.M. '08, Vicedo '09]
Good news
Same situation for Principal Chiral Model, Coset Models !
r and s do not stem from any (a,b,c,d)
First difficulty for the superstring
Bad news
r/s algebra of the continuumcannot be discretized
as a lattice Freidel-Maillet algebra
More difficultiesCould say:
1. Start from known PB algebra of the Lax matrix on continuum
2. Compute then PB of
Problem: These PB are not well defined
Schematically, comes from:
Old problem ! What to do ?
Characteristic function of the interval: Undefined when two points coincide !
Faddeev-Reshetikhin approach [FR '86]
Concerns SU(2) Principal Chiral Model
FR Strategy = To get rid of Non-ultralocality
Satisifies a non-ultralocal r/s algebra
First steps of FR approach
First steps of FR approach
Degeneracy of ultralocal bracket
1. Only possible to reproduce Reduction of PCM dynamics defined by setting Casimirs to constants
2. Can be done in a consistent way because these quantities are chiral/antichiral
Reduction of conformal symmetry
Hamiltonian H' for reduced dynamics
How to generalize FR approach to the superstring ?
- Keep Lax matrix
Impossible to guess !
Try to mimick FR:
But- No sign of Coset
- Its Casimirs are inconsistent with the dynamics of the superstring !
How to generalize FR approach to the superstring ?
The way out of the tunnel
Needs to understand better the algebraic setting behind Hamiltonian integrability of the superstring
and more precisely, the deep origin of Non-ultralocality.
Rephrase integrability in the right framework
= R-matrix approach
Already done by B. Vicedo in 1003.1192
[Semenov-Tian-Shansky '83]
Quartet behind integrability
Quartet behind integrability
Origin of non-ultralocality
Origin of non-ultralocality
First try
Impossible to completely get rid of non-ultralocality !
Good lead: Can recast original FR procedure in this language
Second try
r and s stem from (a,b,c,d) satisfying Freidel-Maillet
conditions !
[Semenov-Tian-Shansky and Sevostyanov '95]
Alleviation of non-ultralocality
r and s stem from (a,b,c,d) satisfying Freidel-Maillet
conditions !
[Semenov-Tian-Shansky and Sevostyanov '95]
Alleviation of non-ultralocalityResulting non-ultralocality is mild
Modified PB for phase space variables
Casimirs of the modified Poisson bracketCase of Sigma Model on symmetric space F/G
Casimirs of the modified Poisson bracketCase of Sigma Model on symmetric space F/G
Casimirs and Pohlmeyer reductionCasimirs are consistent with dynamics of Sigma Model !
[Pohlmeyer '76]
Result of the procedure
Non-ultralocality of symmetric space sine-Gordon models viewed as gauged WZW models + Potential is mild
Pohlmeyer reduction of Sigma model on symmetric space F/G = SSSG model
Consequence for symmetric space sine-Gordon (SSSG) models
[Eichenherr, Pohlmeyer '79]
[Bakas et al. '96,Grigoriev and Tseytlin '07]
Case of the superstring
[Grigoriev and Tseytlin '07, Mikhailov and Schäfer-Nameki '07]
Remarks
Pohlmeyer reduction[Mikhailov '05, '06, Schmiddt '10, '11]
Remarks
[Dorey and Hollowood '95, Hoare and Tseytlin '10]
[Maillet '86]
ConclusionNon-ultralocality of generalized sine-Gordon models is mild
Challenge: Reach same situation as for ultralocal models. One knows from [Freidel-Maillet '91 '92] that one has to search for
representation of Quantum Algebra
Generalization of first steps of FR procedure=
Pohlmeyer reduction
Conclusion
Appealing structure which brings hope that one may be able to quantize from first principles
at least the Pohlmeyer reduction of the superstring !